Stock Deduplication

After years working with major players in a wide array of different industry sectors, ranging from Patent and Trademarks Offices to Law Enforcement agencies or media companies, LTU has learned that large digital photo collections usually contain a significant amount of multiple copies of the same photo, that have been resized, reencoded or edited. This amount can range from less than 1% up to 20%.

Now, for a stock photo agency or a news agency, as well as for any organization storing photos or videos, the consequences of having twice the same photo or video in a collection can be a serious:

• All the costs related to the photo are doubled: the photo was paid for twice, manually annotated twice, and is using twice the amount of storage, bandwidth and backup space it should.
• The photo can be an IP or licensing conflict - what if the same photo was sold to you by to different photographers under two different licensing modes? What if the same photo has different copyright information attached?
• On user generated exchange platforms or microstock sites, abusive users could have uploaded the same photo a number of times to raise the chance it will be viewed, thus cheating on the revenue the photo can generate, at the expense of the rest of your community or users.

LTU engine can be easily integrated into your web server or back office platform to automatically check that any addition to your catalog is not already in the catalog. To do so, you will first need to index all your images with LTU engine, through it's very simple REST API. Once this is complete, LTU engine can be used to compare any new image to the existing repository of images. If a tentative addition already has a derivative or copy within the database, it can be automatically blocked or submitted to a moderation process - the derivative can be acceptable. Also, while indexing the existing catalog, it is very easy to perform the same checks and detect existing duplicates within the current catalog.